Part 2 of a two-part series.
- Steven Seltzer, MD, Radiology Department Chair Emeritus at Brigham and Women’s Hospital and Distinguished Professor at Harvard Medical School
- Andrew Menard, JD, Executive Director of Radiology at Brigham and Women’s Hospital
- Clayton M. Christensen, PhD, Kim B. Clark Professor of Business Administration at Harvard Business School
- Edward Prewitt, Editorial Director for NEJM Catalyst, moderator
Ed Prewitt: The lack of interoperability is a criticism that has been made of the meaningful use funds investment. It’s like the one thing that wasn’t required for EMRs, and yet what I’m taking from this conversation is perhaps that would not have made a difference. Clay, what you’re saying is there’s a requirement of technological maturity, and there’s essentially growth pains that every industry has to go through.
Clayton Christensen: That’s right. Here’s a way to think about it. There are two types of architectures in the world. One type of architecture we would call an interdependent architecture, and that’s defined by, everything you do to one piece has implications for how you do the other pieces of the system. If you do one thing here, you have to do something else there, and that means that you have to be integrated. You have to do everything in order to do anything in that world, which we define as interdependent architecture. And almost always at the beginning of an industry’s history, the architecture has to be interdependent because you can’t articulate how this fits with that. That has to emerge. Nobody is smart enough to know at the beginning exactly how they [should fit].
Steven Seltzer: And that’s where we are now, effectively.
Christensen: That’s right, and people who play that game well are big companies that put their arms around the whole thing, and when they change this, they can say, “All right, change that.” But at the other end of the spectrum — and this is a spectrum — at the other end are architectures that are modular. And when it’s modular, I know exactly if I change this in this component, I have to change that in that component, and when you understand what each piece has to contribute to the whole, then you don’t have to do the whole thing anymore. . . . But [modularity] comes toward the end because you need to be able to articulate what you need, how to measure what you need, and so on.
Prewitt: What is it going to take to move to EMRs that work the way that the buyers would like them to and that everyone intended, and how long is it going to take? Steven?
Seltzer: You know, Ed, we also have billions of dollars in sunk costs, so that for a large health system in an aggressive way to say, “I don’t want this anymore,” that’s going to be a very difficult decision to do. All these things collude, I think, to say it’s going to take a while to work our way out. . . . The current state is pretty fixed for quite some time, and there will be incremental progress, of course. Every year, there’s a new software release, and it’s better, but nothing dramatic until something else changes.
Prewitt: What are the factors, the most important factors, that need to be solved in order to move to a solution?
Andrew Menard: Well, you even have to define solution, I suppose, but if we’re thinking about the word interoperability, if we define solution from the perspective of the patient and the ability to move easily, to have continuity of care and bring their information with them, just that basic idea, I think that — again, curious about my co-panelists’ view — there are some opportunities for outsiders in that case. And there are some ways to use things that go beyond the control of the health systems where patients are more active; patients [become] more IT literate, and there are other tools.
There are companies as big as Apple and many small ones who may be able to offer utilities that assist with some of these problems that look like they’re on the fringe right now, but could lead toward solutions that sort of are almost rebellious — that disrupt from the outside. I don’t see a lot of incentive for a big system, as Steven points out, that has put a billion dollars into architecture, into infrastructure, to self-disrupt, and so I think of it as an outsider insurgency.
Seltzer: That’s what Clay says in his book, that . . . the disruptors will come in with 70% of the capabilities in the existing systems at 1/100 of the cost.
Menard: And even more orthogonal than that, they’re coming at it from a completely different direction.
Seltzer: Correct. They say this interdependent architecture just doesn’t fit in health care, so let’s just go about it in a completely different way.
Christensen: Yeah, almost always, disruption wins by coming at the bottom of the market with simple problems and simple answers to simple problems.
Prewitt: There are all these other competitors out there outside of the big five or so. Do any of you see positive signs today for, say, a web-based EMR or an EMR that does something different that could work to achieve the ends we’re talking about, or is it going to have to be an orthogonal entrant that nobody’s really thinking about today?
Seltzer: Well, there’s talk in every company about moving the data off your mainframe, which is basically what we have now, and putting that in the cloud and having all the interactions be through web-based interfaces to get more flexibility. It’s been 5 years or so, and despite the rhetoric about it, there’s not much that has changed in a large interdependent system. Again, if we wait a little bit longer, that may be possible because . . . technology has advanced in things like cloud storage and the desire to amalgamate patient data from across health systems, in the interest of providing better predictive and preventive care and more precision diagnosis. Those things may put pressure because it’s technologically feasible, and there are medical imperatives to do it. Those might accelerate the development of these smaller APIs and web apps.
Prewitt: NEJM Catalyst’s core audience is health care providers. Our readers are physicians and other clinicians, they’re clinical leaders, they’re health care executives. What should they be doing about the problems that they are experiencing with EMRs? And there are all the problems that we’ve talked about: frustration, cost, burnout, lack of interoperability. Steven?
Seltzer: I think they will have to be patient because the changes won’t come quickly. They need to be proactive about being honest about what the issues are and not just burying them that “this is just a corporate thing and that we’ll never make it any better.” To paraphrase what Clay was saying, be open to considering different practice and payment systems that could obviate some of the complexity of what EHRs do now, which is billing, because if you’re not on a transactional basis any longer — you do need data, clinical data that should be interoperable — but you don’t need to have the ICD-10 and CPT code for everything that you do because you’re being paid a fixed amount. It’s appealing in that it takes a tremendous amount of overhead and complexity out of the system, and ironically, obviates the need for the important kludgy part of the EMR. Those would be things that they could open to.
Menard: Let’s not forget — and I’m not a doctor, but I’m a patient sometimes, and so is everybody I know — [that there are] business decisions where it’s not about optimizing a software system. It’s not about optimizing a financial system, a billing system, a collection system, getting our denials down. We really focus on such things as that. Sometimes denials are OK. And it’s usually a doctor — and I’m going to credit Mass General — they’re very good at standing up and saying, “You’re too focused on denials. You’re too focused on these financial metrics.” Sometimes it’s OK to deliver this care even if we’re not going to get paid. That’s a value system that comes through, a value statement that come through. I think we have to not get too overly into our own science of business.